Articles

Ipatasertib plus paclitaxel versus placebo plus paclitaxel as

first-line therapy for metastatic triple-negative breast cancer
(LOTUS): a multicentre, randomised, double-blind,
placebo-controlled, phase 2 trial
Sung-Bae Kim*, Rebecca Dent*, Seock-Ah Im, Marc Espi, Sibel Blau, Antoinette R Tan, Steven J Isakoff, Mafalda Oliveira, Cristina Saura,
Matthew J Wongchenko, Amy V Kapp, Wai Y Chan, Stina M Singel, Daniel J Maslyar, Jos Baselga, on behalf of the LOTUS investigators

Summary
Lancet Oncol 2017; 18: 136072 Background The oral AKT inhibitor ipatasertib is being investigated in cancers with a high prevalence of PI3K/AKT
Published Online pathway activation, including triple-negative breast cancer. The LOTUS trial investigated the addition of ipatasertib to
August 8, 2017 paclitaxel as first-line therapy for triple-negative breast cancer.
http://dx.doi.org/10.1016/
S1470-2045(17)30450-3
Methods In this randomised, placebo-controlled, double-blind, phase 2 trial, women aged 18 years or older with
See Comment page 1293
measurable, inoperable, locally advanced or metastatic triple-negative breast cancer previously untreated with
*Contributed equally
systemic therapy were recruited from 44 hospitals in South Korea, the USA, France, Spain, Taiwan, Singapore, Italy,
Listed in the appendix
and Belgium. Enrolled patients were randomly assigned (1:1) to receive intravenous paclitaxel 80 mg/m (days 1, 8,
Department of Oncology, Asan
15) with either ipatasertib 400 mg or placebo once per day (days 121) every 28 days until disease progression or
Medical Center, University of
Ulsan College of Medicine, Seoul, unacceptable toxicity. Randomisation was by stratified permuted blocks (block size of four) using an interactive
South Korea (Prof S-B Kim MD); web-response system with three stratification criteria: previous (neo)adjuvant therapy, chemotherapy-free interval,
Division of Medical Oncology, and tumour PTEN status. The co-primary endpoints were progression-free survival in the intention-to-treat population
National Cancer Centre,
and progression-free survival in the PTEN-low (by immunohistochemistry) population. This ongoing trial is registered
Singapore, Singapore
(R Dent MD); Department of with ClinicalTrials.gov (NCT02162719).
Internal Medicine, Seoul
National University Hospital, Findings Between Sept 2, 2014, and Feb 4, 2016, 166 patients were assessed for eligibility and 124 patients were
Cancer Research Institute, Seoul
National University College of
enrolled and randomly assigned to paclitaxel plus ipatasertib (n=62) or paclitaxel plus placebo (n=62). Median
Medicine, Seoul, South Korea follow-up was 104 months (IQR 65141) in the ipatasertib group and 102 months (60136) in the placebo
(S-A Im MD); Breast Disease group. Median progression-free survival in the intention-to-treat population was 62 months (95% CI 3890) with
Center, Hospital Saint Louis, ipatasertib versus 49 months (3654) with placebo (stratified hazard ratio [HR] 060, 95% CI 037098; p=0037)
Paris, France (M Espi MD);
Northwest Medical Specialties
and in the 48 patients with PTEN-low tumours, median progression-free survival was 62 months (95% CI 3691)
and Division of Oncology, with ipatasertib versus 37 months (1973) with placebo (stratified HR 059, 95% CI 026132, p=018). The most
University of Washington, common grade 3 or worse adverse events were diarrhoea (14 [23%] of 61 ipatasertib-treated patients vs none of
Washington, WA, USA 62 placebo-treated patients), neutrophil count decreased (five [8%] vs four [6%]), and neutropenia (six [10%] vs
(S Blau MD); Levine Cancer
Institute, Carolinas HealthCare
one [2%]). No colitis, grade 4 diarrhoea, or treatment-related deaths were reported with ipatasertib. One treatment-
System, Charlotte, NC, USA related death occurred in the placebo group. Serious adverse events were reported in 17 (28%) of 61 patients in the
(A R Tan MD); Massachusetts ipatasertib group and nine (15%) of 62 patients in the placebo group.
General Hospital, Boston, MA,
USA (S J Isakoff MD); Medical
Oncology Department, Vall
Interpretation Progression-free survival was longer in patients who received ipatasertib than in those who received
dHebron University Hospital, placebo. To our knowledge, these are the first results supporting AKT-targeted therapy for triple-negative breast
Vall dHebron Institute of cancer. Ipatasertib warrants further investigation for the treatment of triple-negative breast cancer.
Oncology, Barcelona, Spain
(M Oliveira MD, C Saura MD);
Genentech Inc, South
Funding F Hoffmann-La Roche.
San Francisco, CA, USA
(M J Wongchenko BS, Introduction The PI3K/AKT signalling pathway is often activated in
A V Kapp PhD, W Y Chan PhD,
The PI3K/AKT signalling pathway plays a crucial part in breast cancer, and has attracted interest as a target in triple-
S M Singel MD, D J Maslyar MD);
and Memorial Sloan Kettering carcinogenesis, promoting cell survival and growth.1,2 negative breast cancer.5,6 Large-scale comp rehensive
Cancer Center, New York, NY, AKT is the central node of the PI3K/AKT pathway.3 genomic analyses have characterised the heterogeneous
USA (J Baselga MD) Phosphatidylinositol (3,4,5)-triphosphate, a direct nature of triple-negative breast cancer, including a
Correspondence to: product of PI3K activity, promotes AKT trafficking to the subgroup with genetic activation of the PI3K/AKT pathway
Prof Sung-Bae Kim, Department cell membrane and association with other cell- signalling through activating mutations in PIK3CA or AKT1, and
of Oncology, Asan Medical
Center, University of Ulsan
proteins.4 Full activation of AKT occurs via alterations in PTEN.79 Additionally, approximately half of
College of Medicine, Songpa-gu, phosphorylation at two threonine and serine residues, triple-negative breast cancers have deficient expression of
Seoul 05505, South Korea leading to phosphorylation and regulation of numerous the tumour suppressor PTEN, which is associated with a
sbkim3@amc.seoul.kr
cellular proteins, including mTORC1 and S6 kinase. higher degree of AKT pathway activation.2,10

1360 www.thelancet.com/oncology Vol 18 October 2017


Research in context
Evidence before this study
We searched PubMed to identify publications published
between Jan 1, 2001, and March 31, 2017, that included the
search terms AKT, PI3K, and triple-negative breast
cancer. We also searched PubMed for publications in the
same period describing assessment of ipatasertib using the
terms ipatasertib or GDC-0068. We did not use any
language restrictions in our search. No previous
randomised trials have investigated the targeting of AKT or
PI3K specifically in triple-negative breast cancer. Analyses of
single-arm studies in mesenchymal and metaplastic
triple-negative breast cancer have suggested a more
pronounced response to a combination of an mTOR
inhibitor, bevacizumab, and pegylated liposomal
doxorubicin in patients with PI3K/AKT/mTOR pathway
aberrations. A phase 1 study showed potent inhibition of
Ipatasertib is a highly selective oral ATP-competitive
small-molecule AKT inhibitor.11 In cell line and
xenograft models, ipatasertib showed activity in a
broad range of cancer types, including prostate, breast,
ovarian, colorectal, and non-small-cell lung cancers.11
Sensitivity to ipatasertib tended to be associated with
high phosphorylated AKT levels, PTEN protein loss or
genetic mutations in PTEN, and PIK3CA mutations,
whereas KRAS and BRAF mutations were typically
associated with resistance to ipatasertib.11 As PI3K/
AKT pathway activation is relevant for survival during
periods of mitotic stress,12 the combination of
ipatasertib and taxanes was explored. Preclinical
studies showed synergy between ipatasertib and
taxanes.13 Analysis of on-study tumour biopsy samples
from a phase 1 clinical study showed robust AKT
pathway inhibition by ipatasertib at clinically achievable
doses.14
Based on these findings and its mechanism of action,
ipatasertib is under clinical assessment in cancers with a
high prevalence of PI3K/AKT pathway activation.
A phase 1 study15 of single-agent ipatasertib in
52 pretreated patients with various tumour types,
including breast cancer, showed an acceptable safety
profile (characterised by gastrointestinal effects, asthenia
or fatigue, and rash) and preliminary antitumour activity.
Of note, many patients with disease stabilisation had
PI3K/AKT pathway-activating alterations in their
tumours. In breast cancer, the combination of ipatasertib
(400 mg once daily, days 121) with paclitaxel 90 mg/m
per week (days 1, 8, and 15), repeated every 28 days, was
well tolerated and showed radiographic responses in the
phase 1b PAM4983g study.13
We report results of a randomised phase 2 trial
investigating the addition of ipatasertib to paclitaxel as
first-line therapy for metastatic triple-negative breast
cancer.
www.thelancet.com/oncology Vol 18 October 2017 1361
Articles
See Online for appendix
AKT signalling with ipatasertib, with notable activity in
metastatic breast cancer showing PTEN loss or PIK3CA/AKT
mutations.
Added value of this study
To our knowledge, these are the first prospective trial results
supporting AKT targeting in triple-negative breast cancer.
Prespecified analyses in the population of patients with
PIK3CA/AKT1/PTEN-altered tumours suggest efficacy of
ipatasertib in this population.
Implications of all of the available evidence
Our results support future investigation of ipatasertib plus
paclitaxel in diseases with high prevalence of PI3K/AKT
pathway activation, particularly in patients with
PIK3CA/AKT1/PTEN-altered tumours.
Methods
Study design and participants
LOTUS is a randomised, double-blind, placebo-controlled,
phase 2 trial. Patients were enrolled at 44 hospitals in
South Korea, the USA, France, Spain, Taiwan, Singapore,
Italy, and Belgium (appendix pp 23).
Eligible patients were women aged 18 years or older,
with Eastern Cooperative Oncology Group performance
status 0 or 1, and locally advanced or metastatic triple-
negative breast cancer (defined as <1% tumour cell
expression of oestrogen and progesterone receptors and
negative HER2 status [fluorescence or chromogenic in-
situ hybridisation {FISH/CISH} HER2/CEP17 ratio <20,
or locally assessed immunohistochemistry 0 or 1+{or 2+
but negative by FISH/CISH}]) not amenable to curative
resection. Patients had to have measurable disease
according to Response Evaluation Criteria in Solid
Tumors (RECIST; version 1.1) and adequate haemato
logical, renal, hepatic, and cardiac functions. A formalin-
fixed paraffin-embedded tumour specimen was required
from all patients for central analysis of PTEN expression
before randomisation. The most recently obtained
tumour sample was requested for submission, but a
fresh biopsy sample was not required and primary
tumour samples were acceptable.
Previous systemic therapy for locally advanced or
metastatic disease was not permitted; however, previous
(neo)adjuvant chemotherapy, radiotherapy, or chemoradio
therapy completed at least 6 months before the first dose
was allowed. Patients were ineligible if they had known
brain or spinal cord metastasis, ongoing grade 2 or worse
peripheral neuropathy or grade 2 or worse uncontrolled
or untreated hypercholesterolaemia or hyper tri
glyceridaemia, or active small or large intestine
inflammation (such as Crohns disease or ulcerative colitis).
All patients provided written informed consent before
undergoing any study-specific procedures. Independent
 Articles
institutional review boards at all participating centres
approved the protocol and all study-related documents.
The protocol is available in the appendix.
Randomisation and masking
Eligible patients were randomly assigned (1:1) to either
ipatasertib plus paclitaxel or placebo plus paclitaxel by
investigators using an interactive web-response system
with an allocation sequence generated by Bracket Global
LCC (Reading, UK). Randomisation was by stratified
permuted blocks (block size of four). Randomisation was
stratified by three criteria: previous (neo)adjuvant
chemotherapy (yes vs no), chemotherapy-free interval
(12 vs >12 months vs no previous chemotherapy), and
central tumour PTEN status as assessed by
immunohistochemistry (H score 0 vs 1150 vs >150). In
some cases, patients were randomly assigned before
PTEN status was available; for stratification, these
patients were assigned to the stratum with an H score
more than 150. This approach was adopted because if
patients were otherwise eligible and able to enrol on the
study, we did not consider it ethically acceptable to delay
their first-line treatment while waiting for centrally
assessed PTEN status or if tissue samples were inadequate
for central PTEN analysis. However, for stratified efficacy
analyses, the actual PTEN status (if known) was used.
The stratification factors of previous (neo)adjuvant
chemotherapy and chemotherapy-free interval partly
overlap. However, our intention was to try to balance the
treatment groups in this heterogeneous treatment setting
not only by sensitivity to previous (neo)adjuvant
chemotherapy, but also according to tumour biology
(depending on priming of the PI3K/AKT signalling
pathway by previous chemotherapy) or clinical features of
recurrence or de-novo stage IV disease that could be
differentiated by previous (neo)adjuvant chemotherapy.
Placebo tablets were identical in shape and colour to
the ipatasertib tablets. Investigators, patients, and the
sponsor were masked to treatment assignment.
Procedures
Patients received intravenous paclitaxel 80 mg/m on
days 1, 8, and 15 of each 28-day cycle in combination
with either oral ipatasertib 400 mg/day or placebo,
administered on days 121 of each 28-day cycle. There is
no standard paclitaxel schedule in metastatic breast
cancer. Investigators indicated a strong preference for the
3 weeks on/1 week off schedule of paclitaxel 80 mg/m per
week when the LOTUS trial was designed. This schedule
has been used in previous clinical studies16,17 and maintains
the cumulative dose intensity achieved with 175 mg/m
every 3 weeks (as recommended in the prescribing
information). Treatment was continued until disease
progression, intolerable toxicity, or withdrawal of consent.
Ipatasertib or placebo could be temporarily interrupted for
up to 4 consecutive weeks if patients had toxicity
considered related to the study drug. Diarrhoea was
1362 www.thelancet.com/oncology Vol 18 October 2017
managed with loperamide or according to institutional
guidelines and standard of care, including but not limited
to therapy with diphenoxylate and atropine, codeine,
or octreotide. If symptoms persisted despite adequate
(combination) antidiarrhoeal medications and dose
interruptions, dose reductions were implemented.
Ipatasertib (or placebo) was initially reduced to
300 mg/day, then to 200 mg/day, and was discontinued
permanently at the third appearance of an adverse event
requiring dose reduction. Paclitaxel dose modifications
were implemented according to standard practice or
institutional guidelines. The protocol suggested a
reduction to 65 mg/m at the first reduction and then
permanent discontinuation if toxicity recurred. All
patients who discontinued study therapy were allowed to
receive subsequent anticancer therapy outside the study
protocol. Disease progression that occurred after initiation
of a new anticancer therapy was not collected per protocol;
in such patients, progression-free survival was censored at
the time of the last tumour assessment.
Tumours were assessed every 8 weeks by the
investigators according to RECIST (version 1.1). After
discontinuation of treatment, patients were followed up
every 3 months for survival and subsequent anticancer
therapies. Safety was assessed and graded according to
National Cancer Institute Common Terminology Criteria
for Adverse Events (version 4.0) on an ongoing basis until
the study drug discontinuation visit (or resolution or
stabilisation of ongoing related adverse events).
Laboratory assessments (including haematology, fasting
serum chemistry, coagulation, fasting lipid profile, and
urinalysis) were done within 48 h before each study drug
administration. Patient-reported outcomes (PROs) were
assessed using the European Organisation for Research
and Treatment of Cancer Core Quality of Life
Questionnaire C30 (EORTC QLQ-C30), which includes
30 questions assessing five functional scales, three
symptom scales, and six single items. Questionnaires
were distributed by staff at the site and completed by the
patient before study assessments or drug administration
on day 1 of every cycle, at treatment discontinuation, and
at tumour follow-up. Pharmacokinetic parameters of
ipatasertib were assessed in all patients by sparse plasma
sampling on day 1 of cycle 1 (052 h and 46 h after study
drug administration) and on day 8 of cycle 1 (02 h and
25 h after study drug administration).
At screening, PTEN status was centrally assessed using
antibody clone 138G6 (cat #9559, Cell Signaling
Technology, Leiden, Netherlands; Targos Molecular
Pathology GmbH, Kassel, Germany). Before the primary
analysis, tumour tissue samples were assessed centrally
by additional molecular assays to define the patient
population with PTEN-low tumours (by immuno histo
chemistry; co-primary endpoint) and the patient
population with PI3K/AKT pathway-activated tumours
(secondary endpoint). For the co-primary endpoint,
PTEN-low tumours were defined as those having

immunohistochemistry 0 in at least 50% of tumour cells
using the Ventana immunohistochemistry assay (clone
SP218; Spring Bioscience, Pleasanton, CA, USA). This
assay was used instead of the one used to determine PTEN
status for stratification because it had undergone a greater
degree of technical validation and is being developed as a
potential companion diagnostic assay for ipatasertib. The
classification of PTEN-low tumours also adopted a scoring
method based on quantification of the number of cells
lacking expression, thus providing a more robust scale to
measure the extent of complete loss of PTEN expression.
The FoundationOne next-generation sequencing assay
(Foundation Medicine, Cambridge, MA, USA)18 was used
to identify patients with PI3K/AKT pathway-activated
tumours, defined as the presence of genetic PTEN-
inactivating alterations or PIK3CA/AKT1-activating
mutations (PIK3CA Arg88Gln, Asn345Lys, Cys420Arg,
Glu542X, Glu545X, Gln546X, Met1043Ile, His1047X, or
Gly1049Arg mutations, where X represents any change in
aminoacid residue, or AKT1 Glu17Lys mutations), referred
to hereafter as PIK3CA/AKT1/PTEN-altered tumours.
Outcomes
The co-primary endpoints were investigator-assessed
progression-free survival in the intention-to-treat
population and progression-free survival in the subgroup
of patients with PTEN-low tumours. Progression-free
survival was defined as the interval between randomisation
and the first occurrence of disease progression or death
from any cause within 30 days of the last dose of study
treatment (death on study). As specified in the protocol,
patients who discontinued study treatment without
documented disease progression were censored at the
date of last tumour assessment before initiation of new
anticancer therapy.
Secondary endpoints were investigator-assessed
confirmed objective response (confirmed by a repeat
assessment at least 4 weeks after the criteria for response
are first met), duration of confirmed objective response
(defined as the interval between first observation of a
confirmed objective response and first observation of
disease progression or death on study as assessed by the
investigator), and overall survival in the intention-to-treat
population and patients with PTEN-low tumours; efficacy
(progression-free survival, confirmed objective response
rate, duration of confirmed objective response, and
overall survival) in patients with PI3K/AKT pathway-
activated tumours; and safety (incidence, nature, and
severity of adverse events). Additional objectives included
assessment of pharmacokinetics; PROs for disease-
related and treatment-related symptoms, patient
functioning, and health-related quality of life; and further
exploratory translational research. We also did post-hoc
analyses of the clinical benefit (defined as either an
objective response, or a best overall response of complete
or partial response or stable disease together with a
progression-free survival of 24 weeks or longer).
www.thelancet.com/oncology Vol 18 October 2017 1363
Articles
166 patients assessed for eligibility
42 ineligible*
124 enrolled
124 randomly assigned
62 assigned to ipatasertib plus paclitaxel 62 assigned to placebo plus paclitaxel
1 received no study
treatment
61 received ipatasertib plus paclitaxel 62 received placebo plus paclitaxel
45 discontinued 52 discontinued placebo
ipatasertib 44 disease progression
36 disease progression 1 symptomatic
1 symptomatic deterioration
deterioration 1 death
4 adverse event 1 adverse event
3 patient withdrawal 1 non-compliance
1 physician decision 3 patient withdrawal
1 physician decision
16 treatment ongoing 10 treatment ongoing
62 included in intention-to-treat analysis 62 included in intention-to-treat analysis
Figure 1: Trial profile
ECOG=Eastern Cooperative Oncology Group. LVEF=left ventricular ejection fraction. *The reasons for screen
failure in 42 patients were: not meeting inclusion criteria (two signed informed consent, two ECOG performance
status 1, one locally advanced or metastatic triple-negative breast cancer not amenable to curative resection,
one measurable disease, and six adequate haematological and organ function), meeting exclusion criteria (one
previous therapy for locally advanced or metastatic triple-negative breast cancer; two radiatiotherapy in previous
28 days; one major surgery, open biopsy, or significant traumatic injury in preceding 30 days; ten known brain or
spinal cord metastasis; one New York Heart Association class II, III, or IV heart failure or LVEF <50%, or active
ventricular arrhythmia requiring medication; one ongoing unstable angina or history of myocardial infarction in
previous 6 months; one grade 3 uncontrolled or untreated hypercholesterolaemia or hypertriglyceridaemia;
three congenital long QT syndrome or screening corrected QT interval 480 ms; three inability to comply with
study and follow-up procedures; one other malignancy within 5 years; three potential contraindication); and
12 other reasons (more than one answer possible). Five patients in the ipatasertib group and six in the placebo
group received new anticancer therapy after discontinuing study therapy before disease progression. Further
details of patients who discontinued without progression and received new anticancer therapy are provided in
the appendix.
Statistical analysis
The planned sample size was 60 patients per group for a
total of 120 patients overall to ensure 83 progression-free
survival events for the primary analysis. As this
hypothesis-generating trial was designed to assess safety
and provide preliminary evidence of activity, it was not
powered to detect minimal clinically meaningful
differences between treatment groups at a significant
level of 5%. Instead, 90% CIs for the hazard ratio (HR)
were calculated, anticipating that for clinically
 Articles

Intention-to-treat population PTEN-low population PIK3CA/AKT1/PTEN-altered

tumour population

Ipatasertib Placebo Ipatasertib Placebo Ipatasertib Placebo

plus paclitaxel plus paclitaxel plus paclitaxel plus paclitaxel plus paclitaxel plus paclitaxel
(n=62) (n=62) (n=25) (n=23) (n=26) (n=16)

Median age (years) 54 (4463) 53 (4563) 50 (4463) 56 (4665) 52 (4463) 53 (4660)

Age group
1840 years 10 (16%) 5 (8%) 4 (16%) 2 (9%) 5 (19%) 1 (6%)
4164 years 40 (65%) 46 (74%) 18 (72%) 15 (65%) 18 (69%) 14 (88%)
65 years 12 (19%) 11 (18%) 3 (12%) 6 (26%) 3 (12%) 1 (6%)
Race
Asian 28 (45%) 30 (48%) 10 (40%) 9 (39%) 16 (62%) 7 (44%)
White 26 (42%) 28 (45%) 12 (48%) 13 (57%) 8 (31%) 9 (56%)
Black or African-American 5 (8%) 3 (5%) 2 (8%) 1 (4%) 0 0
Other 3 (5%) 1 (2%) 1 (4%) 0 2 (8%) 0
ECOG performance status
0 44 (71%) 36 (58%) 17 (68%) 15 (65%) 13 (50%) 9 (56%)
1 18 (29%) 22 (35%) 8 (32%) 7 (30%) 13 (50%) 7 (44%)
Missing 0 4 (6%) 0 1 (4%) 0 0
Previous (neo)adjuvant 41 (66%) 40 (65%) 19 (76%) 15 (65%) 18 (69%) 10 (63%)
chemotherapy*
Anthracycline 34 (55%) 34 (55%) 16 (64%) 12 (52%) 14 (54%) 7 (44%)
Taxane 31 (50%) 34 (55%) 18 (72%) 14 (61%) 12 (46%) 7 (44%)
Chemotherapy-free interval (months)*
12 18 (29%) 16 (26%) 8 (32%) 4 (17%) 7 (27%) 3 (19%)
>12 23 (37%) 24 (39%) 11 (44%) 11 (48%) 11 (42%) 7 (44%)
None 21 (34%) 22 (35%) 6 (24%) 8 (35%) 8 (31%) 6 (38%)
PTEN H score*
0 10 (16%) 11 (18%) 9 (36%) 7 (30%) 8 (31%) 3 (19%)
1150 27 (44%) 27 (44%) 10 (40%) 12 (52%) 6 (23%) 6 (38%)
>150 25 (40%) 24 (39%) 6 (24%) 4 (17%) 12 (46%) 7 (44%)
Histopathological subtype
Ductal 59 (95%) 59 (95%) 25 (100%) 23 (100%) 24 (92%) 15 (94%)
Lobular 3 (5%) 1 (2%) 1 (4%) 0 2 (8%) 0
Tubular 1 (2%) 3 (5%) 0 0 1 (4%) 1 (6%)
Metastatic sites
Lung 27 (44%) 32 (52%) 13 (52%) 14 (61%) 13 (50%) 9 (56%)
Liver 19 (31%) 17 (27%) 7 (28%) 6 (26%) 7 (27%) 5 (31%)
Lymph nodes 36 (58%) 38 (61%) 14 (56%) 18 (78%) 15 (58%) 12 (75%)
Bone 16 (26%) 17 (27%) 7 (28%) 10 (43%) 5 (19%) 8 (50%)

Data are n (%) or median (IQR). ECOG=Eastern Cooperative Oncology Group. IXRS=interactive web-response system. *Stratification factor, reported per IXRS. Data not from
IXRS. More than one answer possible.

Table 1: Baseline characteristics

meaningful outcomes, the upper limit of the 90% CI
would be less than 1. We report 95% CIs to be consistent
with published literature. The primary analysis was
intended to include 50 progression-free survival events
in patients with PTEN-low tumours. Assuming
60% prevalence of PTEN-low tumours, we anticipated
83 progression-free survival events in the intention-to-
treat population.
Efficacy analyses were based on all randomly assigned
patients (intention-to-treat population). Analyses for
the co-primary endpoints were stratified; the Cox
proportional hazard model included the treatment group
and three stratification factors as covariates. In this proof-
of-concept study, the definition of progression-free
survival for the primary endpoint was chosen with the
aim of identifying antitumour activity closely related to

1364 www.thelancet.com/oncology Vol 18 October 2017

 Articles

Diagnostic
prevalence (%)*

PTEN low by IHC 48

PTEN altered by NGS 17
PIK3CA/AKT1 mutant by NGS 25
Primary or metastatic
n=86 evaluable by IHC and NGS
n=101 evaluable by Ventana IHC
n=103 evaluable by FMI NGS

Figure 2: Biomarker prevalence

IHC=immunohistochemistry. NGS=next-generation sequencing. FMI=Foundation Medicine Inc. *Prevalence based on all available diagnostic data. Each vertical set of blocks represents an individual
patients tumour. Green blocks represent PTEN loss by IHC; pink blocks represent PTEN-altered by NGS; dark blue blocks represent PIK3CA/AKT1-mutant by NGS; grey blocks represent samples with no
corresponding data available (assay failure or insufficient sample for testing). The bottom row shows whether samples are from primary (light blue) or metastatic (red) tumour sites.

study treatment. The risk of bias was reduced by the

Ipatasertib plus paclitaxel (n=61) Placebo plus paclitaxel (n=62)
double-blinded, placebo-controlled trial design. However,
recognising that in a poor-prognosis disease setting such Treatment duration (months)
as triple-negative breast cancer, this definition might lead Ipatasertib or placebo 50 (3578) 35 (1654)
to censoring of events related to disease progression, Paclitaxel 41 (3272) 35 (1551)
progression-free survival including death from any cause Cumulative dose intensity*
irrespective of time from last dose was included as a Ipatasertib or placebo 990 (9171000) 1000 (9451000)
sensitivity analysis, otherwise using the same approach Paclitaxel 1000 (9091000) 1000 (9381000)
as for the primary analysis. All other analyses were not
Mean (SD) cumulative dose intensity*
stratified; the only covariate in Cox proportional hazard
Ipatasertib or placebo 933 (110) 955 (119)
models was the treatment group.
Safety analyses were based on all patients who received Paclitaxel 942 (118) 955 (113)
at least one dose of ipatasertib, placebo, or paclitaxel; Treatment discontinued for adverse event
patients were analysed based on the treatment actually Ipatasertib or placebo 4 (7%) 1 (2%)
received. PRO analyses were done on all patients in the Paclitaxel 5 (8%) 5 (8%)
intention-to-treat population with a baseline assessment Treatment interrupted for adverse event
and at least one post-baseline assessment. The full Ipatasertib or placebo 22 (36%) 12 (19%)
intention-to-treat set was used to assess compliance and
Paclitaxel 31 (51%) 30 (48%)
completion rates, summarised at each timepoint by
Dose reduced for adverse event
treatment group with reasons for missing data. Summary
Ipatasertib or placebo 13 (21%) 4 (6%)
statistics of linear transformed scores were reported for
all scales of the EORTC QLQ-C30 according to the Paclitaxel 23 (38%) 7 (11%)
EORTC scoring manual guidelines for each assessment Data are n (%) or median (IQR) unless otherwise stated. *Cumulative dose intensity for each study drug (ipatasertib,
timepoint. The mean change of the linear transformed placebo, or paclitaxel) was calculated using the actual amount of study drug received in mg divided by the expected
scores from baseline (and 95% CI using the normal amount of study drug in mg. The expected amount of study drug was calculated based on treatment duration (the
interval between the first and last administered doses of study drug) and the initial dose and schedule specified in the
approximation) were reported. Changes from baseline of protocol. Diarrhoea, asthenia, and vomiting (n=1), diarrhoea (n=1), tuberculosis (n=1), and embolism (n=1).
10 points or more in PRO scores were defined as clinically Cholestasis or cell death. Hypoaesthesia (n=1), pharyngitis or tonsillitis (n=1), tuberculosis (n=1), back pain (n=1),
meaningful.19 Efficacy, safety, and PRO analyses were not and embolism (n=1). Peripheral sensory neuropathy (n=2), neuropathy peripheral (n=1), cholestasis/cell death (n=1),
neutrophil count decreased (n=1).
adjusted for multiple comparisons.
Cumulative dose intensity for each study drug was Table 2: Treatment exposure (safety population)
calculated using the actual amount of study drug
received in mg divided by the expected amount of study
drug in mg. The expected amount of study drug was serum albumin, liver function tests, and serum
calculated based on treatment duration (the interval creatinine were tested for significance on pharmaco
between the first and last administered doses of study kinetic parameters of interest.
drug) and the initial dose and schedule specified in the Analyses were done using SAS (version 9.4).
protocol. An internal monitoring committee reviewed partly
Ipatasertib plasma concentration versus time data unblinded summaries of the safety data approximately
were pooled and analysed using a population-based 16 weeks after enrolment of the first 20 patients. After
pharmacokinetic (popPK) modelling approach. Non- completion of enrolment and occurrence of approximately
linear mixed-effect modelling was used for the 40 progression-free survival events, the internal
estimation of popPK parameters for ipatasertib. monitoring committee reviewed an interim safety and
Covariates such as patient demographics, total protein, efficacy analysis.

www.thelancet.com/oncology Vol 18 October 2017 1365

 Articles

the study, were involved in writing the report, and

A
100 Ipatasertib plus paclitaxel (n=62)
approved the final version for submission. The first and
Placebo plus paclitaxel (n=62) second authors had final responsibility for the decision
to submit for publication.
Progression-free survival (%)

80

60 Results
Between Sept 2, 2014, and Feb 4, 2016, 166 patients were
40
assessed for eligibility and 124 patients were randomly
20
assigned to treatment with ipatasertib (62 patients) or
Stratified HR 060 (95% CI 037098) placebo (62 patients; figure 1). One randomly assigned
Log-rank p=0037
0 patient who received no study treatment was excluded
0 2 4 6 8 10 12 14 16 18 from the safety analysis population. Baseline character
Number at risk
(number censored)
istics were generally balanced between treatment groups
Ipatasertib plus paclitaxel 62 50 (4) 31 (9) 22 (13) 14 (15) 11 (15) 6 (19) 2 (21) 1 (22) 0 (23) (table 1). Biomarker-assessable populations for both
Placebo plus paclitaxel 62 43 (3) 23 (12) 13 (12) 10 (12) 6 (13) 3 (14) 0 (17) PTEN status and PIK3CA/AKT1/PTEN alterations
B showed similar baseline characteristics to the intention-
100 Ipatasertib plus paclitaxel (n=25) to-treat population. In 11 patients (four randomly assigned
Placebo plus paclitaxel (n=23) to placebo and seven randomly assigned to ipatasertib),
the PTEN status used for stratification differed from that
Progression-free survival (%)

80
used for analysis. Of these, five (two placebo, three
60 ipatasertib) were classified as having a PTEN status of
more than 150 at randomisation but their PTEN status in
40
analyses was unknown, four (one placebo, three
20 ipatasertib) were classified as having a PTEN status of
Stratified HR 059 (95% CI 026132)
Log-rank p=018
more than 150 at randomisation but in the analyses, their
0 PTEN H score was 1150, and one (placebo) was classified
0 2 4 6 8 10 12 14 16 18
with a PTEN status of 0 at randomisation but was
Number at risk
(number censored) classified with a PTEN H score of 1150 in the analyses.
Ipatasertib plus paclitaxel 25 21 (0) 12 (4) 9 (5) 5 (6) 4 (6) 0 (9) The remaining patient (ipatasertib) was classified as
Placebo plus paclitaxel 23 15 (1) 8 (3) 6 (3) 5 (3) 3 (3) 2 (3) 0 (5)
having a PTEN status of more than 150 at randomisation
C but in the analyses her PTEN H score was 0.
100 Ipatasertib plus paclitaxel (n=26) Samples were centrally assessable for PTEN in
Placebo plus paclitaxel (n=16) 101 (81%) of 124 patients; in the remaining 23 (19%)
Progression-free survival (%)

80 patients, PTEN status could not be determined because

of assay failure or insufficient sample for testing. Of
60 these 101 assessable samples, 48 (48%) were classified as
PTEN low, lower than the 60% predicted. Of the
40
103 patient samples assessed by next-generation
20 sequencing, 42 (41%) had PIK3CA/AKT1/PTEN-altered
Non-stratified HR 044 (95% CI 020099) tumours (appendix p 4). Of the 15 patients with PTEN
0 genetic alterations by next-generation sequencing, and
0 2 4 6 8 10 12 14 16 18
samples assessed for immunohistochemistry, 14 (93%)
Time (months)
Number at risk had loss of PTEN protein expression (figure 2). However,
(number censored)
Ipatasertib plus paclitaxel 26 22 (3) 13 (7) 10 (9) 7 (10) 5 (10) 3 (12) 1 (13) 1 (13) 0 (14)
a substantial proportion of patients with loss of PTEN
Placebo plus paclitaxel 16 11 (1) 7 (2) 4 (2) 3 (2) 2 (2) 1 (2) 0 (3) protein expression did not have a genetic alteration
(figure 2). Of the 21 patients with activating mutations in
Figure 3: Progression-free survival in the (A) intention-to-treat population, (B) PTEN-low subgroup, PIK3CA and AKT1 and samples assessed for
(C) PIK3CA/AKT1/PTEN-altered subgroup.
HR=hazard ratio. immunohistochemistry, only six (29%) had PTEN loss by
immuno histochemistry (figure 2). Prevalence of
This trial is registered with ClinicalTrials.gov, number PIK3CA/AKT1/PTEN alterations did not differ between
NCT02162719. primary (n=76) and metastatic (n=27) samples (figure 2)
or between samples that were collected before
Role of the funding source administration of (neo)adjuvant chemotherapy and
The funder of the study was involved in the study design, samples collected from chemotherapy-naive patients
data collection, data analysis, data interpretation, and (data not shown).
writing of the report, and gave approval to submit it for Treatment exposure is summarised in table 2. At the
publication. All authors had full access to all the data in clinical cutoff date (June 7, 2016), the median duration of

1366 www.thelancet.com/oncology Vol 18 October 2017

 Articles

Number of events/patients Median progression-free survival Non-stratified

(95% CI), months hazard ratio
(95% Wald CI)
Ipatasertib plus Placebo plus Ipatasertib plus Placebo plus
paclitaxel paclitaxel paclitaxel paclitaxel

Age (years)
<50 16/22 17/24 62 (3691) 29 (1854) 062 (031124)
50 23/40 28/38 57 (37129) 51 (3763) 062 (035109)
DFI since last chemotherapy (months)
12 8/11 12/14 44 (1973) 35 (1639) 046 (016136)
>12 16/31 20/28 91 (67NA) 54 (3673) 049 (025095)
No prior chemotherapy 15/20 13/20 37 (3690) 54 (2891) 100 (046217)
(Neo)adjuvant chemotherapy
Yes 24/42 31/41 72 (5593) 37 (2951) 048 (028081)
No 15/20 14/21 37 (3690) 54 (2891) 103 (048220)
Targos IHC PTEN status (H-score)
0 5/11 7/10 72 (55NA) 18 (18109) 021 (006075)
1150 23/30 22/29 37 (3357) 51 (3673) 101 (056183)
>150 9/18 15/21 90 (55NA) 37 (1963) 036 (015086)
Unknown 2/3 1/2 72 (1672) 54 (NA) 082 (0051324)
Ventana IHC PTEN status (staining intensity)
PTEN low 16/25 18/23 62 (3691) 36 (1973) 068 (035135)
PTEN non-low 17/29 19/24 67 (37129) 50 (2955) 054 (027107)
Unknown 6/8 8/15 57 (3393) 51 (2954) 051 (016163)
NGS PIK3CA/AKT1/PTEN
Altered 12/26 13/16 90 (46NA) 49 (3663) 044 (020099)
Non-altered 21/28 23/33 53 (3673) 36 (2955) 076 (041140)
Unknown 6/8 9/13 57 (3273) 51 (17110) 083 (029242)
All 39/62 45/62 62 (3890) 49 (3654) 062 (040095)

02 05 1 2 4

Favours ipatasertib Favours placebo

plus paclitaxel plus paclitaxel

Figure 4: Subgroup analysis of progression-free survival

Non-stratified analysis. DFI=disease-free interval. IHC=immunohistochemistry. NA=not assessible. NGS=next-generation sequencing.

follow-up was 104 months (IQR 65141) in the
ipatasertib group and 102 months (60136) in the
placebo group. Among patients who discontinued study
treatment before disease progression, the proportions
subsequently receiving non-study systemic anticancer
therapy were similar in the two treatment groups (five [8%]
of 62 patients in the ipatasertib group and six [10%] of
62 patients in the placebo group). These patients were
censored at the date of last tumour assessment before
initiation of new anticancer therapy except for two patients
(both with a progression-free survival event; appendix p 5).
The primary progression-free survival analysis was
triggered by reaching approximately 83 progression-free
survival events in the intention-to-treat population:
39 events in the ipatasertib group and 45 in the placebo
group. One patient in each group died without evidence
of progression; the remaining events were disease
progression. Median progression-free survival was
62 months (95% CI 3890) with ipatasertib versus
49 months (3654) with placebo (stratified HR 060,
95% CI 037098; log-rank p=0037; figure 3A). In the
sensitivity analysis, including all deaths from any cause,
the stratified HR was 066 (95% CI 041106; log-rank
p=0081); median progression-free survival
was 59 months (95% CI 3873) with ipatasertib versus
50 months (3654) with placebo.
At the time of data cutoff, progression-free survival
events had been documented in 34 (71%) of 48 patients in
the PTEN-low population (16 [64%] of 25 in the ipatasertib
group and 18 [78%] of 23 in the placebo group). In this
population, median progression-free survival was
62 months (95% CI 3691) with ipatasertib versus
37 months (1973) with placebo (stratified HR 059,
95% CI 026132, log-rank p=018; figure 3B).
Prespecified analyses in the subgroup of 42 patients
with PIK3CA/AKT1/PTEN-altered tumours, after
progression-free survival events in 12 (46%) of 26 patients
in the ipatasertib group and 13 (81%) of 16 patients in the
placebo group, showed median progression-free survival
of 90 months (95% CI 46not assessable) with
ipatasertib versus 49 months (3663) with placebo
(non-stratified HR 044, 95% CI 020099, log-rank
p=0041; figure 3C). In patients with PIK3CA/AKT1/
PTEN-non-altered tumours, with progression-free
survival events in 21 (75%) of 28 patients in the ipatasertib
group and 23 (70%) of 33 patients in the placebo group,

www.thelancet.com/oncology Vol 18 October 2017 1367

 Articles
Intention-to-treat population
Ipatasertib Placebo
plus paclitaxel plus paclitaxel
(n=62) (n=62)
Objective response 25 (40%) 20 (32%)
Duration of response (months) 79 (56NA) 74 (3992)
Clinical benefit 30 (48%) 23 (37%)
Data are n (%) or median (95% CI). Objective response is per Response Evaluation Criteria in Solid Tumors version 1.1. Clinical benefit is defined as either an objective
response, or a best overall response of complete or partial response or stable disease together with a progression-free survival of 24 weeks. NA=not assessable.
Table 3: Summary of secondary efficacy endpoints
the median progression-free survival was 53 months
(95% CI 3673) in the ipatasertib group versus
37 months (2955) in the placebo group (non-stratified
HR 076, 95% CI 041140, p=036). Progression-free
survival in selected subgroups, including the
randomisation stratification factors, is shown in figure 4.
Further anticancer therapy was administered after
disease progression to 30 (77%) of 39 patients in the
ipatasertib group and 38 (84%) of 45 patients in the
placebo group whose disease had progressed by the time
of data cutoff. Overall survival results are immature, with
deaths in nine (15%) of 62 patients in the ipatasertib
group and 17 (27%) of 62 patients in the placebo group.
The primary cause of death was disease progression in
22 (85%) of 26 patients. Secondary endpoints of objective
response and duration of response, and the post-hoc
assessment of clinical benefit are shown in table 3.
Median duration of response was similar in the two
treatment groups for the intention-to-treat and PTEN-
low populations, but was longer in ipatasertib-treated
patients compared with placebo in the PIK3CA/AKT1/
PTEN-altered subgroup.
The most common adverse events of any grade in the
ipatasertib group were gastrointestinal effects
(diarrhoea, nausea, and vomiting), alopecia, neuropathy,
fatigue, and rash (table 4). These were typically grade 1
or 2 in severity. Grade 3 or worse adverse events
occurred in 33 (54%) of 61 patients in the ipatasertib
group and 26 (42%) of 62 patients in the placebo group
(table 4). The most common individual grade 3 or worse
adverse events in the ipatasertib group were diarrhoea,
neutrophil count decreased, and neutropenia). The
most common grade 3 or worse adverse events by
grouped term (appendix p 6) were diarrhoea (14 [23%]
of 61 ipatasertib-treated patients vs none of 62 placebo-
treated patients), neutropenia (comprising neutropenia,
neutrophil count decreased, and febrile neutropenia;
11 [18%] vs five [8%]), peripheral neuropathy (comprising
peripheral sensory neuropathy, neuropathy peripheral,
paraesthesia, hypo aesthesia, dysaesthesia, muscular
weakness, neuro toxicity, and peripheral motor
neuropathy; four [7%] vs three [5%]), fatigue or asthenia
(three [5%] vs four [6%]), and pneumonia or lower
1368 www.thelancet.com/oncology Vol 18 October 2017
PTEN-low subgroup by PIK3CA/AKT1/PTEN-altered subgroup
Immunohistochemistry by next-generation sequencing
Ipatasertib Placebo Ipatasertib Placebo
plus paclitaxel plus paclitaxel plus paclitaxel plus paclitaxel
(n=25) (n=23) (n=26) (n=16)
12 (48%) 6 (26%) 13 (50%) 7 (44%)
65 (44NA) 75 (73NA) 112 (56NA) 61 (3876)
14 (56%) 7 (30%) 14 (54%) 7 (44%)
respiratory tract infection (three [5%] vs none). Of note,
there were no episodes of grade 4 diarrhoea and no
reported cases of colitis. Serious adverse events were
more common in the ipatasertib group (17 [28%] of
61 patients, predominantly infections and gastro
intestinal effects) than in the placebo group (nine [15%]
of 62 patients, predominantly infections). Four patients
had adverse events resulting in death: one patient with
pneumonia in the ipatasertib group (not considered
related to study treatment) and three in the placebo
group (one case each of cholestasis together with cell
death [reported by the investigator as cytolysis (liver),
both events assessed as related to placebo and
paclitaxel]; metastatic breast cancer; and death from
unknown cause 287 days after the last dose of the study
drug).
Diarrhoea typically occurred during the first cycle of
ipatasertib (median time to onset 5 days) but some
cases were observed in later cycles. Diarrhoea led to
discontinuation of ipatasertib in two (3%) of 61 patients,
dose reduction of ipatasertib in eight (13%), and
temporary interruption of ipatasertib in four (7%). Anti-
diarrhoeal drugs (predominantly loperamide) were
administered in 39 (64%) of 61 patients in the ipatasertib
group and six (10%) of 62 patients in the placebo group.
At data cutoff, almost all episodes of diarrhoea had
resolved.
There was high compliance with PRO assessment
questionnaires: more than 90% of patients in each
treatment group completed at least one item of the
EORTC QLQ-C30 at each cycle (appendix p 7). In both
treatment groups, mean change from baseline scores for
most functional scales (cognitive, physical, and social)
and the global health status/quality-of-life domain were
not clinically meaningful according to the predefined
threshold of a 10-point change from baseline (appendix
pp 89). Similarly, no clinically meaningful changes were
observed for most of the disease-related and treatment-
related symptom scales (appetite loss, constipation,
dyspnoea, nausea and vomiting, insomnia, pain, and
financial difficulties) up to and including cycle 5.
However, in the ipatasertib group, a clinically meaningful
improvement in emotional functioning was observed at
 Articles

Ipatasertib plus paclitaxel (n=61) Placebo plus paclitaxel (n=62)

All grades Grade 3 Grade 3* Grade 4* All grades Grade 3 Grade 3* Grade 4*

All 61 (100%) 33 (54%) 27 (44%) 5 (8%) 60 (97%) 26 (42%) 20 (32%) 3 (5%)

Diarrhoea 57 (93%) 14 (23%) 14 (23%) 0 12 (19%) 0 0 0
Alopecia 33 (54%) 0 0 0 29 (47%) 0 0 0
Nausea 30 (49%) 1 (2%) 1 (2%) 0 21 (34%) 1 (2%) 1 (2%) 0
Vomiting 17 (28%) 2 (3%) 2 (3%) 0 14 (23%) 0 0 0
Peripheral sensory neuropathy 16 (26%) 3 (5%) 3 (5%) 0 10 (16%) 2 (3%) 2 (3%) 0
Fatigue 16 (26%) 2 (3%) 2 (3%) 0 21 (34%) 4 (6%) 4 (6%) 0
Rash 16 (26%) 1 (2%) 1 (2%) 0 12 (19%) 0 0 0
Asthenia 15 (25%) 2 (3%) 2 (3%) 0 6 (10%) 0 0 0
Myalgia 15 (25%) 0 0 0 15 (24%) 0 0 0
Neutropenia 13 (21%) 6 (10%) 4 (7%) 2 (3%) 15 (24%) 1 (2%) 0 1 (2%)
Decreased appetite 13 (21%) 0 0 0 11 (18%) 0 0 0
Stomatitis 11 (18%) 1 (2%) 1 (2%) 0 5 (8%) 0 0 0
Constipation 11 (18%) 0 0 0 10 (16%) 0 0 0
Dizziness 11 (18%) 0 0 0 9 (15%) 0 0 0
Insomnia 11 (18%) 0 0 0 8 (13%) 0 0 0
Neuropathy peripheral 10 (16%) 0 0 0 14 (23%) 1 (2%) 1 (2%) 0
Dermatitis acneiform 10 (16%) 0 0 0 5 (8%) 0 0 0
Neutrophil count decreased 9 (15%) 5 (8%) 3 (5%) 2 (3%) 9 (15%) 4 (6%) 3 (5%) 1 (2%)
Headache 9 (15%) 1 (2%) 1 (2%) 0 12 (19%) 0 0 0
Abdominal pain 9 (15%) 0 0 0 7 (11%) 0 0 0
Pyrexia 9 (15%) 0 0 0 6 (10%) 0 0 0
Arthralgia 9 (15%) 0 0 0 6 (10%) 0 0 0
Anaemia 8 (13%) 2 (3%) 2 (3%) 0 8 (13%) 2 (3%) 2 (3%) 0
Dyspepsia 8 (13%) 0 0 0 6 (10%) 0 0 0
Pruritus 8 (13%) 0 0 0 5 (8%) 0 0 0
Nasopharyngitis 8 (13%) 0 0 0 5 (8%) 0 0 0
Cough 7 (11%) 0 0 0 8 (13%) 0 0 0
Pneumonia 3 (5%) 3 (5%) 2 (3%) 0 0 0 0 0
Febrile neutropenia 3 (5%) 2 (3%) 1 (2%) 1 (2%) 0 0 0 0
Hypertension 3 (5%) 1 (2%) 1 (2%) 0 3 (5%) 2 (3%) 2 (3%) 0
Bone pain 3 (5%) 0 0 0 2 (3%) 1 (2%) 0 1 (2%)
Hypocalcaemia 2 (3%) 1 (2%) 0 1 (2%) 0 0 0 0
Thrombocytopenia 1 (2%) 0 0 0 1 (2%) 1 (2%) 0 1 (2%)
Cholestasis 1 (2%) 0 0 0 1 (2%) 1 (2%) 0 0
Death 0 0 0 0 1 (2%) 1 (2%) 0 0
Pancytopenia 0 0 0 0 1 (2%) 1 (2%) 0 1 (2%)
Cell death 0 0 0 0 1 (2%) 1 (2%) 0 0
Breast cancer metastatic 0 0 0 0 1 (2%) 1 (2%) 0 0
Fever 0 0 0 0 1 (2%) 1 (2%) 0 1 (2%)

Adverse events in 10% or more of patients (any grade), all grade 3 or worse in more than one patient in either treatment group, or any grade 4 or 5. *Worst grade
reported (eg, a patient who has an event at both grade 3 and grade 4 appears only in the grade 4 column). Grade 5 in one patient (2%). Unmapped (verbatim term
shown).

Table 4: Summary of adverse events in the safety population

cycle 2, whereas a clinically meaningful worsening was after cycle 5 are not described due to sample size attrition
observed for diarrhoea (cycles 25), fatigue (cycle 5), and (in both groups, fewer than 50% of patients remained on
role functioning (cycles 35). Scores from timepoints treatment beyond cycle 5).

www.thelancet.com/oncology Vol 18 October 2017 1369

 Articles
The plasma concentrations of ipatasertib obtained by
sparse sampling in this study were consistent with known
pharmacokinetic profiles and overall characteristics of
ipatasertib and its metabolite G-037720. Exploratory
analyses showed no relationship between ipatasertib
exposure and incidence of diarrhoea, neutropenia, and
neuropathy (data not shown).
Discussion
Results of the randomised, double-blind, placebo-
controlled, phase 2 LOTUS trial show that adding
ipatasertib to paclitaxel as first-line therapy for triple-
negative breast cancer increased progression-free survival
compared with that for placebo plus paclitaxel. The
increase in median progression-free survival was quite
modest in the intention-to-treat population and PTEN-
low sub group but more pronounced in predefined
analyses of the patient population with
PIK3CA/AKT1/PTEN-altered tumours characterised by
next-generation sequencing. Overall, adverse events were
consistent with previous experience, manageable, and
reversible.
The 49-month median progression-free survival in
the control group of the intention-to-treat population
was within the range reported in subgroup analyses
of patients with triple-negative breast cancer in previous
randomised trials (46 months in the E2100 trial,20
55 months in the NU07B1 trial,21 and 63 months in the
MERiDiAN22 trial). Of note, among patients in LOTUS
who had previously received (neo)adjuvant chemotherapy,
approximately a third had disease recurrence within
12 months of chemotherapy (25 [30%] of the 84 patients
who had received prior chemotherapy), whereas such
patients were excluded from the MERiDiAN trial.22
Similarly, the 32% of patients achieving response in the
control group of LOTUS seems to be consistent with the
available data reported in the literature (21% with a
higher starting dose of single-agent paclitaxel in a mixed
population of triple-negative breast cancer and non-
triple-negative breast cancer patients in the E2100 trial;20
28% with paclitaxel plus onartuzumab in a randomised
phase 2 trial in triple-negative breast cancer
predominantly in the first-line setting23).
When the trial was designed, it was anticipated that
patients with PTEN-low tumours by immunohisto-
chemistry might derive increased benefit from ipatasertib.
This was hypothesised because a randomised phase 2 trial
in metastatic castration-resistant prostate cancer showed
that the effect of ipatasertib was more pronounced in
the subgroup of patients with PTEN loss identified by
immunohistochemistry or next-generation sequencing.24
However, PTEN loss is only one of several mechanisms
leading to activation of the PI3K/AKT pathway. In breast
cancer, activating mutations in PIK3CA and AKT1 are
frequently observed, whereas in castration-resistant
prostate cancer, these mutations are rare.8,9,25 In our study
population, a substantial proportion of patients with
1370 www.thelancet.com/oncology Vol 18 October 2017
PTEN-low tumours by immunohistochemistry did not
have a genetic alteration. This is consistent with previous
reports of non-genetic loss of PTEN in triple-negative
breast cancer.26 In the LOTUS trial, the effect of ipatasertib
in the subgroup of patients with PTEN-low tumours by
immunohistochemistry was no greater than in those with
PTEN-non-low tumours or in the intention-to-treat
population. However, efficacy analysis in the population
with PIK3CA/AKT1/PTEN-altered tumours supporting
the studys secondary objectives showed an encouraging
progression-free survival HR of 044 and an increase of
41 months in the median progression-free survival
(median 90 months in the ipatasertib group vs 49 months
in the placebo group). Duration of response results
supported these findings. This difference in efficacy based
on absence of expression of PTEN through non-genetic
mechanisms compared with loss of PTEN function
through mutations and copy-number loss could be a key
difference in how PTEN loss might drive tumours and be
PI3K/AKT-addicted in prostate versus breast cancers.
The most common adverse events were gastrointestinal,
in particular diarrhoea. Most cases of diarrhoea were
grade 1 or 2; grade 3 diarrhoea occurred in 23% of patients
and there were no grade 4 cases. Diarrhoea was manageable
and reversible, and only two patients discontinued
ipatasertib because of diarrhoea. Of note, primary
prophylactic antidiarrhoeal drugs were not specified as
part of safety management guidelines in the protocol.
Although patients in the ipatasertib plus paclitaxel
group had clinically meaningful worsening in patient-
reported role function, diarrhoea, and fatigue, patients
overall global health status or quality of life was
maintained up to and including cycle 5. There was also
no clinically meaningful change in cognitive, physical, or
social function scales or other symptom scales. Together,
our results indicate that the tolerability of the ipatasertib
plus paclitaxel regimen might allow rational combination
with other carefully selected agents.
The similar pharmacokinetic profiles in this study and
previous experience suggest that there was no paclitaxel
ipatasertib drug interaction that affected metabolism or
clearance (as predicted from the phase 1b study and
preliminary assessment by population pharmacokinetic
methodology [Roche data on file]). Although exploratory
analyses showed no clear relationship between ipatasertib
exposure and incidence of diarrhoea, neutropenia,
and neuropathy, assessment of the exposureresponse
relationship is difficult in a trial with only one dose level.
One of the main limitations of these results is the small
sample size. The biomarker-selected population showing
the most encouraging effect of ipatasertib includes only
42 patients; despite ours being prespecified analyses, our
findings should be interpreted with caution and require
prospective validation. Furthermore, although baseline
characteristics in this population were generally
balanced, randomisation was not stratified by next-
generation sequencing results.

The prevalence of PIK3CA/AKT1/PTEN alterations
was 41%. We observed no clinically significant difference
in the prevalence of PIK3CA/AKT1/PTEN alterations in
primary versus metastatic samples. We also saw no
difference in alteration frequency between samples that
were collected before administration of (neo)adjuvant
chemotherapy and samples collected from chemotherapy-
naive patients. However, this analysis does not preclude
the possibility that these alterations might be enriched in
patients with metastatic disease. The apparent absence of
treatment effect in the population of patients who had
received no previous chemotherapy (most of whom had
de-novo stage IV disease at study entry) should be
interpreted with caution due to the small sample size.
There have been few targeted therapy advances in
the management of triple-negative breast cancer;
chemotherapy (with or without the anti-angiogenic
agent bevacizumab) remains the standard of care for
these patients, who typically have a poor prognosis and
no targeted treatment options. Randomised phase 3
trials specifically in triple-negative breast cancer
have reported median progression-free survival of
approximately 35 months and median overall survival
of approximately 12 months with chemotherapy
alone.27,28 Ipatasertib in combination with paclitaxel is
one of several novel strategies under assessment in
randomised trials in triple-negative breast cancer29 and
the treatment landscape could change substantially in
the near future. If emerging agents fulfil their potential,
treatment decision making and sequencing could
become increasingly complex, and biomarker selection
is expected to play an important part in individualised
therapy for the heterogeneous collection of diseases
traditionally described as triple-negative breast cancer.
Our findings warrant further prospective investigation
of ipatasertib in the population of patients with
PIK3CA/AKT1/PTEN-altered tumours. Additional
research in triple-negative breast cancer includes the
randomised phase 2 FAIRLANE trial (NCT02301988),
which is assessing the addition of ipatasertib to paclitaxel
in the neoadjuvant setting.30 Results from FAIRLANE
might provide further information on patient selection,
although, as in LOTUS, patients were not stratified by
PIK3CA/AKT1/PTEN-altered tumours.
Although the development of ipatasertib to date has
focused on triple-negative breast cancer, Lin and
colleagues11 observed similar sensitivity to ipatasertib in
HER2-positive and hormone receptor-positive cell lines.
Our results support future assessment of ipatasertib plus
paclitaxel in diseases with high prevalence of PI3K/AKT
pathway activation and in particular in patients with
PIK3CA/AKT1/PTEN-altered tumours. A phase 3 trial in
metastatic breast cancer is underway.
Contributors
S-BK, RD, SJI, MO, and CS were involved in the design of the study.
S-BK, RD, S-AI, ME, SB, ART, SJI, MO, CS, and JB obtained the data.
MJW, AVK, WYC, SMS, and DJM analysed the data. All authors
www.thelancet.com/oncology Vol 18 October 2017 1371
Articles
interpreted the data, reviewed and revised the draft manuscripts, and
approved the final version for submission.
Declaration of interests
S-BK reports research funding from Novartis, Sanofi Aventis, Kyowa Kirin
Inc, and Dongkook Pharma Co, Ltd. RD reports honoraria for consultancy,
advisory boards, and speaker engagements from Pfizer, Roche, Eisai,
Merck, Novartis, and AstraZeneca and research funding from
AstraZeneca. S-AI reports honoraria for advisory roles from Novartis,
Hanmi, and Spectrum and a research grant from AstraZeneca. ME reports
honoraria from Pfizer and Merck Sharp & Dohme, and research funding
to his institution from Roche and Novartis. SB reports a contract (but no
honoraria) for consulting and advisory roles for BMS and declares that her
husband receives royalties associated with his position as a professor at the
University of Washington and is the owner of the company All4cure. ART
reports honoraria for an advisory board from AbbVie, and research
funding (to institution) from Merck, Pfizer, and Genentech. SJI reports
clinical research funding from Genentech. MO reports honoraria for
consulting and advisory roles from Genentech/Roche. CS reports
honoraria for consulting and advisory roles from Puma Biotechnology,
Pfizer, and Roche and research funding (to her institution) from
Genentech and AstraZeneca. MJW is an employee of Genentech, Inc, and
holds shares in Roche and Ariad Pharmaceuticals. AVK, WYC, SMS, and
DJM are employees of Genentech, Inc, and hold stock in Roche. JB reports
no competing interests.
Acknowledgments
This study was funded by F Hoffmann-La Roche. We thank all the
patients who agreed to take part in the trial. We also thank the
investigators who participated and the Genentech, Roche, and Array
BioPharma study teams. We thank the members of the internal
monitoring committee, the study clinical operations team, the
biostatisticians, Premal Patel (trial design and initiation), and
Agnes Hong (PRO interpretation). Medical writing assistance was
provided by Jennifer Kelly, (Medi-Kelsey Ltd, Ashbourne, UK), funded by
F Hoffmann-La Roche.
References
1 Cantley LC. The phosphoinositide 3-kinase pathway. Science 2002;
296: 165557.
2 LoRusso PM. Inhibition of the PI3K/AKT/mTOR pathway in solid
tumors. J Clin Oncol 2016; 34: 380315.
3 Manning BD, Cantley LC. AKT/PKB signaling: navigating
downstream. Cell 2007; 129: 126174.
4 Bhaskar PT, Hay N. The two TORCs and Akt. Dev Cell 2007;
12: 487502.
5 Juric D, Castel P, Griffith M, et al. Convergent loss of PTEN leads to
clinical resistance to a PI(3)K inhibitor. Nature 2015; 518: 24044.
6 Di Cosimo S, Baselga J. Management of breast cancer with targeted
agents: importance of heterogeneity. Nat Rev Clin Oncol 2010;
7: 13947.
7 Cancer Genome Atlas Network. Comprehensive molecular portraits
of human breast tumours. Nature 2012; 490: 6170.
8 Curtis C, Shah SP, Chin SF, et al. The genomic and transcriptomic
architecture of 2000 breast tumours reveals novel subgroups.
Nature 2012; 486: 34652.
9 Pereira B, Chin SF, Rueda OM, et al. The somatic mutation profiles
of 2433 breast cancers refines their genomic and transcriptomic
landscapes. Nat Commun 2016; 7: 11479.
10 Millis SZ, Gatalica Z, Winkler J, et al. Predictive biomarker
profiling of >6000 breast cancer patients shows heterogeneity in
TNBC, with treatment implications. Clin Breast Cancer 2015;
15: 47381.
11 Lin J, Sampath D, Nannini MA, et al. Targeting activated Akt with
GDC-0068, a novel selective Akt inhibitor that is efficacious in
multiple tumor models. Clin Cancer Res 2013; 19: 176072.
12 Skladanowski A, Bozko P, Sabisz M, Larsen AK. Dual inhibition of
PI3K/Akt signaling and the DNA damage checkpoint in
p53-deficient cells with strong survival signaling: implications for
cancer therapy. Cell Cycle 2007; 6: 226875.
13 Isakoff SJ, Infante JR, Juric J, et al. Phase Ib dose-escalation study
of the Akt inhibitor ipatasertib (Ipat) with paclitaxel (P) in patients
(pts) with advanced solid tumors. Ann Oncol 2014;
25 (suppl 4): iv148.
 Articles
14 Yan Y, Serra V, Prudkin L, et al. Evaluation and clinical analyses of
downstream targets of the Akt inhibitor GDC-0068. Clin Cancer Res
2013; 19: 697686.
15 Saura C, Roda D, Rosell S, et al. A first-in-human phase I study of
the ATP-competitive AKT inhibitor ipatasertib demonstrates robust
and safe targeting of AKT in patients with solid tumors.
Cancer Discov 2017; 7: 10213.
16 Sato K, Inoue K, Saito T, et al. Multicenter phase II trial of weekly
paclitaxel for advanced or metastatic breast cancer: the Saitama
Breast Cancer Clinical Study Group (SBCCSG-01). Jpn J Clin Oncol
2003; 33: 37176.
17 Miles D, Kim S-B, McNally VA, et al. COLET (NCT02322814):
A multistage, phase 2 study evaluating the safety and efficacy of
cobimetinib (C) in combination with paclitaxel (P) as first-line
treatment for patients (pts) with metastatic triple-negative breast
cancer (TNBC). Proc AmSoc Clin Oncol 2016;
34 (suppl): abstr TPS1100.
18 Frampton GM, Fichtenholtz A, Otto GA, et al. Development and
validation of a clinical cancer genomic profiling test based on
massively parallel DNA sequencing. Nat Biotechnol 2013;
31: 102331.
19 Osoba D, Rodrigues G, Myles J, Zee B, Pater J. Interpreting the
significance of changes in health-related quality of life scores.
J Clin Oncol 1998; 16: 13944.
20 Miller K, Wang M, Gralow J, et al. Paclitaxel plus bevacizumab
versus paclitaxel alone for metastatic breast cancer. N Engl J Med
2007; 357: 266676.
21 Gradishar WJ, Kaklamani V, Sahoo TP, et al. A double-blind,
randomised, placebo-controlled, phase 2b study evaluating
sorafenib in combination with paclitaxel as a first-line therapy in
patients with HER2-negative advanced breast cancer. Eur J Cancer
2013; 49: 31222.
22 Miles D, Cameron D, Bondarenko I, et al. Bevacizumab plus
paclitaxel versus placebo plus paclitaxel as first-line therapy for
HER2-negative metastatic breast cancer (MERiDiAN):
a double-blind placebo-controlled randomised phase III trial with
prospective biomarker evaluation. Eur J Cancer 2017; 70: 14655.
1372 www.thelancet.com/oncology Vol 18 October 2017
23 Diras V, Campone M, Yardley DA, et al. Randomized, phase II,
placebo-controlled trial of onartuzumab and/or bevacizumab in
combination with weekly paclitaxel in patients with metastatic
triple-negative breast cancer. Ann Oncol 2015; 26: 190410.
24 De Bono JS, De Giorgi U, Massard C, et al. PTEN loss as a
predictive biomarker for the AKT inhibitor ipatasertib combined
with abiraterone acetate in patients with metastatic castration-
resistant prostate cancer (mCRPC). Ann Oncol 2016;
27 (suppl 6): 718O.
25 Robinson D, Van Allen EM, Wu YM, et al. Integrative clinical
genomics of advanced prostate cancer. Cell 2015; 161: 121528.
26 Khan S, Kumagai T, Vora J, et al. PTEN promoter is methylated in a
proportion of invasive breast cancers. Int J Cancer 2004; 112: 40710.
27 Tutt A, Ellis P, Kilburn L, et al. The TNT trial: a randomized
phase III trial of carboplatin (C) compared with docetaxel (D) for
patients with metastatic or recurrent locally advanced triple negative
or BRCA1/2 breast cancer (CRUK/07/012). Cancer Res 2015;
75 (suppl 9): abstr S3-01.
28 OShaughnessy J, Schwartzberg L, Danso MA, et al. Phase III study
of iniparib plus gemcitabine and carboplatin versus gemcitabine
and carboplatin in patients with metastatic triple-negative breast
cancer. J Clin Oncol 2014; 32: 384047.
29 Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L.
Triple-negative breast cancer: challenges and opportunities of a
heterogeneous disease. Nat Rev Clin Oncol 2016; 13: 67490.
30 Isakoff SJ, Saura C, Calvo I, et al. FAIRLANE: a phase II
randomized, double-blind, study of the Akt inhibitor ipatasertib
(GDC-0068) in combination with paclitaxel as neoadjuvant
treatment for early stage triple-negative breast cancer.
Proc AmSoc Clin Oncol 2016; 34 (suppl): abstr TPS1105.